Presentation by TTI researcher, Mohammad Poorsartep, at the 2016 Combined Accident Reduction Effort (CARE) Conference, September 26-28 in San Antonio, Texas.

1. Implications of Automated
Vehicle Crash Scenarios
Mohammad Poorsartep
Research Scientist
m-poorsartep@tti.tamu.edu

2. Topics
• Brief History
• Levels and taxonomy
• How does it all work
• How does it “not” work
• Current regulatory environment
• Q&A

3. History of AV

4. SAE Levels
Source: Society of Automotive Eng. /Ricardo

5. How does it work?
• How does it work?
Strategic
Route
Planning
Tactical
Sense
Analyze
Decide
Operational
Brake-Accel. Steer

6. How does it work?
We have a brain for one reason and one
reason only — and that’s to produce
adaptable and complex movements
- Daniel Wolpert, Univ. of Cambridge

7. How does it work?
Source: Automotive-eetimes.com

8. How does it work?
Source: Google

9. Source: Mandli

10. How does it work?
Source: Mandli

11. How does it work?
Pavement
Messages
Bike Lanes
Intersections
Drainage
Islands
Edge Type
Signal
Cabinet
Power
Pedestal
Signal Pole
Manholes
Source: Mandli

12. How does it work?
Fused data / mile
Approx. 4m miles of roads in US
12MB/mile = ~82TB
# of Features Size
Raw Data - 110GB
Map after feature extraction ~6m 4.7GB
Map after tracking and VocabID ~400k 18MB
Map after Subsampling ~180k 12MB

13. How does it “not” work?
Strategic
Route
Planning
Tactical
Sense
analyze
Decide
Operational
Brake-Accel. Steer

14. How does it “not” work?
• Possible Failure Modes:
1. Ghost objects
2. Electro/mechanical components failure
3. wrong information
4. Unexpected events
5. False driver input/action
6. …….

15. How does it “not” work?

16. How does it “not” work?
Source: Washington Post

17. Current Regulatory
Environment
Source: Stanford Univ.

18. Current Regulatory
Environment: NV
For purposes of this chapter, unless the context otherwise requires, a person
shall be deemed the “operator” of an autonomous vehicle which is operated in
autonomous mode when the person causes the autonomous vehicle to
engage, regardless of whether the person is physically present in the vehicle
while it is engaged.
“Autonomous technology” means technology which is installed on a motor
vehicle and which has the capability to drive the motor vehicle without the
active control or monitoring of a human operator. The term does not include an
active safety system or a system for driver assistance, including, without
limitation, a system to provide electronic blind spot detection, crash avoidance,
emergency braking, parking assistance, adaptive cruise control, lane keeping
assistance, lane departure warning, or traffic jam and queuing assistance,
unless any such system, alone or in combination with any other system,
enables the vehicle on which the system is installed to be driven without the
active control or monitoring of a human operator.
“Autonomous vehicle” means a motor vehicle that is equipped with
autonomous technology.

19. Current Regulatory
Environment: CA
An “operator” of an autonomous vehicle is the person who is seated in
the driver’s seat, or if there is no person in the driver’s seat, causes the
autonomous technology to engage.
“Autonomous technology” means technology that has the capability
to drive a vehicle without the active physical control or monitoring by a
human operator.
“Autonomous vehicle” means any vehicle equipped with autonomous
technology that has been integrated into that vehicle.
– An autonomous vehicle does not include a vehicle that is equipped with
one or more collision avoidance systems, including, but not limited to,
electronic blind spot assistance, automated emergency braking
systems, park assist, adaptive cruise control, lane keep assist, lane
departure warning, traffic jam and queuing assist, or other similar
systems that enhance safety or provide driver assistance, but are not
capable, collectively or singularly, of driving t

20. Crash Scenarios
Source: 9to5Google

22. Crash Scenario-1
Operator in the loop
• Driver/operator is in the control loop is expected to take over
of the driving task, if prompted. (SAE Levels 2 and 3)
• An accident is caused due to:
1. Operators failure to monitor the driving environment and taking proper
actions
2. Operator fails to take over the driving task
3. Operator fails to properly engage the system and unable to provide
correcting countermeasures
4. System fails to accurately monitor and detect events in surrounding
environment
5. System fails to act upon the information received.

23. Crash Scenario-1
Operator in the loop
• Level-3 examples (link)

24. Crash Scenario-1
Operator in the loop
Operator fails to
properly engage
the system and
unable to
provide
correcting
countermeasures
(link)

25. Crash Scenario-1
Operator in the loop
System fails to
accurately monitor
and detect events
in surrounding
environment
(link)

26. Crash Scenario-2
Operator NOT in the loop
• Driver/operator is NOT in the control loop and
the system is expected to perform all tasks
related to driving task. (SAE Levels 4 and 5)
• An accident is caused due to:
1. System fails to accurately monitor and detect events in
surrounding environment
2. System fails to act upon the information received.

27. Crash Scenario-2
Operator NOT in the loop
• Level – 4 example (link)

28. Crash Scenario-2
Operator NOT in the loop
• Level – 5 example (link)

29. Crash Scenario-2
Operator NOT in the loop
System fails to
accurately
monitor and
detect events
in surrounding
environment

30. int use_track = 0, use_rects = 1;
// if (t->vmag > 4)
// use_rects = 0;
if (t->vmag > 3.0 && t->maturity > 8)
use_track = 1;
double MAX_DIM = 10;
if (t->box.size[0] > MAX_DIM || t->box.size[1]
> MAX_DIM)
use_track = 0;
Source: John Leanord, MIT

31. Available Data
• Google car rear-ended (link)

32. Resources
• http://d2dtl5nnlpfr0r.cloudfront.net/tti.tamu.edu/documents/PRC-15-46-F.pdf

33. Resources
https://www.transportation.gov/sites/dot.gov/files/docs/AV%20policy%20guidance%20PDF.pdf